JOURNAL OF

POLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPER

Charge Transport and Structure in Semimetallic Polymers

Sam Rudd,1 Juan F. Franco-Gonzalez,2 Sandeep Kumar Singh,2 Zia Ullah Khan ,2
Xavier Crispin ,2 Jens W. Andreasen,3 Igor Zozoulenko ,2 Drew Evans 1
1
Thin Film Coatings Group, Future Industries Institute, University of South Australia, Mawson Lakes, South Australia 5095,
Australia
2
Department of Science and Technology, Organic Electronics, Linkoping University, Norrkoping SE-601 74, Sweden
3
Department of Energy Conversion and Storage, Frederiksborgvej 399, Technical University of Denmark, Roskilde 4000, Denmark
Correspondence to: D. Evans (E-mail: drew.evans@unisa.edu.au)

Received 20 April 2017; accepted 3 October 2017; published online 16 October 2017
DOI: 10.1002/polb.24530

ABSTRACT: Owing to changes in their chemistry and structure,
polymers can be fabricated to demonstrate vastly different
electrical conductivities over many orders of magnitude. At the
high end of conductivity is the class of conducting polymers,
which are ideal candidates for many applications in low-cost
electronics. Here, we report the influence of the nature of the
Incidence Wide Angle X-ray Scattering, Density Functional The-
ory calculations, and Molecular Dynamics simulations indicate
that the chosen doping anion modifies the way PEDOT chains
stack together. This link between structure and specific anion
doping at high doping levels has ramifications for the fabrica-
tion of conducting polymer-based devices. V C 2017 The Authors.

doping anion at high doping levels within the semi-metallic
conducting polymer poly(3,4-ethylenedioxythiophene) (PEDOT)
on its electronic transport properties. Hall effect measurements
on a variety of PEDOT samples show that the choice of doping
anion can lead to an order of magnitude enhancement in
the charge carrier mobility > 3 cm2/Vs at conductivities
approaching 3000 S/cm under ambient conditions. Grazing
Journal of Polymer Science Part B: Polymer Physics Published
by Wiley Periodicals, Inc. J. Polym. Sci., Part B: Polym. Phys.
2018, 56, 97–104
KEYWORDS: charge transport; conducting polymers; DFT; DFT
calculations; GIWAXS; MD simulations; molecular dynamics;
WAXS

1 2 preparation of PEDOT doped with polystyrenesulfonate

INTRODUCTION Synthesis of polyacetylene and polyaniline
has sparked the interest of many researchers over the past
four decades to understand the charge transport mecha-
nism(s) within conducting polymers. Factors on many differ-
ent length scales within the polymeric material are considered
important to the transport of charge, in turn then defining the
conductivity of the polymer; chemical structure, oxidation
level, conjugation length, and morphology, to name but a few.
Despite ongoing studies to elicit new knowledge, the common
understanding is that the properties of conducting polymers
are dominated by the transport of holes (positive charges)
along the conjugated network or “backbone” along the length
of the polymer chain. This understanding leads to the classifi-
cation of conducting polymers as p-type materials.
Through variation of the precursor monomer, fabrication
process, and doping anion, many conducting polymers have
been studied, such as polyaniline,3 polypyrrole,4 and
poly(3,4-ethylenedioxythiophene) (PEDOT).5 Chemical
(PSS–) has yielded conductivity of over 2000 S/cm by post-
treatment (adding an ionic liquid6 or treating with an acid7
are two examples). Further increases in the conductivity to
3400 S/cm was achieved through use of (i) tosylate (Tos–)
as the doping anion, (ii) amphiphilic copolymers, and (iii)
the vapor phase polymerization (VPP) process.8 Adding to
this, Gueye et al. used the post-treatment process with sev-
eral solvents to enhance the Tos– plus amphiphilic copolymer
system yielding PEDOT conductivities of greater than
5000 S/cm.9 Thiophene-based polymers have been employed
in organic photovoltaics and other field effect devices owing
to appreciable charge carrier mobility.10 The study of Cho
et al.11 further pushed the electrical properties of PEDOT, by
creating nanowires of single crystal PEDOT doped with Cl–
via VPP having electrical conductivity exceeding 7600 S/cm
with a concomitant charge carrier mobility approaching
88 cm2/Vs. In more detailed studies by Bubnova et al.,12 the
differences from the use of PSS– or Tos– were shown to arise

Additional Supporting Information may be found in the online version of this article.
C 2017 The Authors. Journal of Polymer Science Part B: Polymer Physics Published by Wiley Periodicals, Inc.
V

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which
permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no
modifications or adaptations are made.

JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104 97


FULL PAPER
from a transition in the conducting polymer from a Fermi
glass to a semimetal respectively. Despite the demonstration
of PEDOT in electrical applications from thermoelectronics13
to spintronics14 to energy storage,15,16 no rationale has been
provided to explain the role of the doping anion in achieving
high electrical conductivity. Notably theoretical modelling,
that is an essential and standard tool in many fields of mate-
rial science, is to a large extend missing in conducting poly-
mer research where the interpretation of experiments
seldom relies on theoretical calculations. Recently, some of
the present authors reported molecular dynamics (MD) sim-
ulations of morphology and crystallization of PEDOT with
tosylate (Tos–) as the doping anion.17 At the same time, the
theoretical understanding of the effect of different anions on
electronic, structural, and morphological properties of the
system at hand remains elusive and practically unexplored.
In this study, we report the electronic properties of conduct-
ing polymer PEDOT with respect to the doping anions using
various experimental techniques [Hall measurements, photo-
absorption, grazing incidence wide-angle X-ray scattering, X-
ray Photoelectron Spectroscopy (XPS), and THz spectros-
copy] combined with the theoretical modeling using atomis-
tic MD simulations of material’s morphology and the density
functional theory (DFT) for calculations of electronic proper-
ties. Importantly this study focuses on high doping levels in
excess of 50% (i.e., 1 anion for 2 EDOT monomers16), where
the conducting polymer (in this case PEDOT) is said to be
semimetallic.12 This is in contrast to the majority of existing
studies which are limited to the doping levels of 33% or less
corresponding to pristine (i.e., as polymerized) PEDOT. A
combination of the experimental characterization and theo-
retical simulations allowed us to outline the origin of the
pronounced difference in transport properties of PEDOT
with different anions, relating them to the differences in the
morphology, which PEDOT exhibits when tosylate is
substituted by different counterions.
EXPERIMENTAL
Sample Preparation
PEDOT was fabricated via VPP using Fe(Tos)3 as the oxidant
in a butanol/ethanol solvent mix, with and without added
PEG-PPG-PEG amphiphilic triblock copolymer (2900 or 5800
Da Mw) (see Supporting Information). Chemically prepared
PEDOT was prepared using polystyrene sulfonate or Tos– as
the doping anion, which was then spin-cast onto substrates.
Samples of the VPP PEDOT:Tos were then electrochemically
reduced or oxidised by placing in a salt solution of Na(NO3),
Li(ClO4), Na2(PhenylPO4), NaCl, and Na(Tos), using the con-
ducting polymer as the working electrode and a Pt wire as
the counter electrode. A reducing potential of 21 V and oxi-
dation potential of 11 V were applied to the samples.
Analysis
Hall effect measurements were made (Ecopus HMS-5300)
using the van der Pauw configuration. Doping level of the
anions were determined using a SPECS (SAGE, Phoibos 150-
98 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104
JOURNAL OF
WWW.POLYMERPHYSICS.ORG POLYMER SCIENCE
HSA) X-ray photoelectron spectroscope (XPS) fitted with a
non-monochromated Al anode, power 200 W, with a base
pressure of 2 3 1026 Pa. For all anions except Tos, the sur-
vey scan was employed to determine the atomic percentage
of the central atom of the anion (N, P, Cl). For all samples
the atomic percentage of S was separated into S in Tos and S
in EDOT by interpreting the S 2p fine scan. This is based on
the peaks attributed to S 2p 1=2 and S 2p 3=2 in Tos and S 2p
1=2 and S 2p 3=2 in PEDOT (the two signals at 167–170 eV
are from Tos, whereas those at 164–167 eV are from
PEDOT). The doping level of anions per EDOT repeat unit
(or number of EDOT repeat units per 1 anion) is calculated
by summation of the atomic percentage of anions relative to
S in EDOT. The full width at half maximum (FWHM) is 1.8
eV for all peaks fitted in the S 2p fine scans, except in the
case of reduced PEDOT (1.5 eV) and when ClO–4 is the sec-
ondary anion (1.6 eV).
Absorption spectra of PEDOT films coated over a glass sub-
strate were obtained using an Agilent Technologies, Cary
5000 series UV–vis-NIR spectrometer. THz spectroscopy was
conducted using the custom-built system described in ref.
18. Grazing Incidence Wide-Angle X-ray Scattering (GIWAXS)
were performed with a setup19 having a rotating Cu-anode
as source, focused by a 1D multilayer mirror providing
monochromatic X-ray radiation (Cu K-alpha, 1.5418 Å) at a
grazing incident angle of 0.188. The scattered radiation was
collected with a Fuji imaging plate.
MD Simulations
MD Simulations were performed using General AMBER Force
Field (GAFF)20 employing the moltemplate code21 in
LAMMPS software suite.22 Water molecules were described
by a model of SPC/E.23 50 PEDOT chains with the oxidation
levels corresponding to those measured by XPS as reported
in Figure 1(c) were used in the calculations. We considered
the chain length of PEDOT N 5 12. Note that the chain length
of PEDOT is not known exactly experimentally but is esti-
mated to be in the range of N 5 10–20 monomer units
depending on the synthesis method employed.24 It is also
believed that short chains are obtained in polythiophenes by
VPP according to the description by Yokozawa et al.25 Addi-
tionally, a narrow and monodispersed molecular weight dis-
tribution is expected, which justifies the use of all of PEDOT
chains with the same chain length in our simulation box.
Finally, we note that that for the case of PEDOT:TOS the cal-
culated morphology was shown to be rather insensitive to
the chain length N.17
The corresponding number of Tos and anions to balance the
charges of the system were considered in a proper propor-
tion as measured by XPS (see Supporting Information Table
S1 for the number of Tos and anios used). Then, the box is
solvated with 30,000 water molecules. All the molecules
were randomly placed in a computational box 12 3 12 3
12 nm3, which typically contained 101,000 atoms. The sys-
tem was then minimized and equilibrated by 20 ns run of
canonical nVT (at 293.15 K) ensemble using the Nose-
Hoover thermostat26 and the time integration method of
 JOURNAL OF
POLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPER

FIGURE 1 The influence of doping anion on VPP PEDOT properties. The (a) absolute electrical conductivity at room temperature,
(b) normalized conductivity as a function of temperature, and (c) anion doping level for different anions in VPP PEDOT. (d) Ratio
of the optical absorption at 1200 nm compared with 850 nm. Determination of (e) the charge carrier density and (f) the charge car-
rier mobility as a function of the electrical conductivity of the PEDOT variant from the Hall effect measurements and XPS. The
open circle in (e) is an electrochemically reduced variant of VPP PEDOT. The open triangles in (f) are mobility and conductivity val-
ues determined from THz reflectance spectroscopy (optical measurement). For (b), (e), and (f) The PEDOT variants are initially
doped with Tos– and subsequently inserted with Tos– (blue) ClO–4 (purple), NO–3 (red), Cl– (green), and PhenylPO422 (black). [Color
figure can be viewed at wileyonlinelibrary.com]

Verlet.27 Then, water was consecutively removed in 7 steps,
such as the water concentration was reduced approximately,
from 82% w.t. (initial solution) to 70, 60, 43, 25, 16, 12, and
finally 0% w.t. (i.e., a dry phase). The system was equili-
brated in each step by a npT (at 1 atm and 293.15 K)
ensemble for 10 ns run with both barostat and thermostat
as Nose-Hoover with corresponding adjustment (decreasing)
of its volume. Also, at each step, as a standard protocol, the
simulations were performed until the potential energy of the
system reached saturation. X-ray diffraction patterns were
simulated as described by Coleman et al.28 and implemented
in LAMMPS suite22 (see the Supporting Information for more
details). Partial charges on each atom of PEDOT and Tos mol-
ecules were calculated using first-principles DFT functional
WB97XD29 with the 6–31 1 g(d) basis set30 as implemented
in Gaussian 09, revision E.01 2009. The partial charge per
atom were taken from the fitting to electrostatic potential
population analysis as implemented in Gaussian suite.31 Fur-
ther details of the employed computational method can be
found in ref. 17.
RESULTS
Charge transport in conducting polymers has been discussed
with respect to holes (positive charges), and how they are
created, stabilized, and traverse through the polymeric mate-
rial.32 Addition of counter-ions (anions) into the polymeric
material, referred to as doping, allows for stabilisation of the
hole through local charge neutralisation (attractive Coulomb
force), thus increasing the lifetime of this charged state.1
Electrochemical oxidation and reduction represents an effi-
cient way to dope or de-dope the polymer by inserting or

JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104 99


FULL PAPER WWW.POLYMERPHYSICS.ORG
removing anions from within the structure. In the case of
PEDOT, insertion (removal) of anions leads to a lowering
(increase) of visible light absorption (ca. 650 nm) with a
concomitant increase (decrease) in the absorption of wave-
lengths above 750 nm. These changes in optical properties
are observed for both polymeric chains in solution and thin
film coatings thereof.33 The visible absorption is associated
with the “neutral” PEDOT chain having no holes along the
conjugated chain. Insertion of anions to create holes leads to
polarons being formed (PEDOT11, absorption ca. 791 and
1300 nm), with further insertion eventually leading to bipo-
larons (PEDOT12, absorption ca. 1170 nm; Supporting Infor-
mation Figure S1). At high oxidation states, where a
bipolaronic network is formed, the bipolaron and valence
band begin to overlap, leading to the classification of the
polymer as semimetallic in nature.12,34
Contrasting views have arisen in the literature about how
the specific chemistry of the doping anion may or may not
influence the charge transport. For example, (i) the doping
anion (Tos– vs. Cl–) yielding no significant changes in electri-
cal properties,35 (ii) the doping anion (PSS– vs. Tos– vs. ClO–4 )
yielding orders of magnitude change in electrical proper-
ties,36 (iii) to the replacement of PSS– with Tos– yielding a
transition from semiconducting to semimetallic behavior.12
To address these differences we herein demonstrate through
experiment and calculation/simulation that transport proper-
ties in PEDOT may be attributed to the doping anion
influencing the structure of the resultant PEDOT.
A variety of PEDOT samples were prepared, using VPP with
the doping anion Tos– and amphiphilic copolymer poly(ethyl-
ene glycol)-poly(propylene glycol)-poly(ethylene glycol) (PEG-
PPG-PEG), and subsequently using ion-exchange to insert dif-
ferent anions. The ion-exchange process of the VPP PEDOT:Tos
samples was achieved by subjecting samples to electrochemi-
cal reduction/oxidation cycling in an aqueous electrolyte solu-
tion of different anions (Tos–, ClO–4 , NO–3 , Cl–, PhenylPO22
4 ) to
move anions in and out of the PEDOT, with a final oxidation
step to insert as many of the anions as possible.
Figure 1 displays the electrical, optical and chemical analysis
of the VPP PEDOT samples. Hall effect measurements were
used to determine the electrical conductivity (r, S/cm) for
the respective samples, including the conductivity as a func-
tion of temperature. Combined with this is the optical (UV–
vis-NIR) and chemical (XPS) analysis of the samples. Firstly,
the chemical analysis revealed that the PEDOT samples had
mixed doping, with both the original Tos– from the VPP pro-
cess and the new anion from the ion-exchange present in the
PEDOT. Given the large reservoir of anions in the electrolyte
for ion exchange, it was envisaged that the anion in solution
would become the main dopant in the PEDOT. However, Tos–
appears to be well bound within the VPP PEDOT.
Comparison of VPP PEDOT with different anions inserted
within the structure shows that the type of the inserted
anion plays an important role on the resultant PEDOT prop-
erties [depicted in Figure 1(a) with respect to the electrical
100 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104
JOURNAL OF
POLYMER SCIENCE
conductivity]. Firstly, the highest electrical conductivity
across the samples is when ClO–4 is employed, which is in
agreement with several studies in the literature for chemi-
cally prepared and electropolymerised PEDOT.36,37 Examina-
tion of the electrical conductivity as a function of
temperature in Figure 1(b) shows all samples display metal-
lic transport properties, similar to that shown for semimetal-
lic PEDOT.12 Inserting PhenylPO22 4 into the PEDOT yields an
electrical conductivity one third that of the PEDOT:Tos refer-
ence sample. In part this is due to the PhenylPO–4 not doping
to the same level as Tos– (1 anion in 2.4 EDOTs vs. 1 in 1.8),
as shown in Figure 1(c) (determined from XPS, see Support-
ing Information Figure S2). It is important to note that the
very high doping level herein compared with that generally
reported as the theoretical limit for PEDOT (1 anion in 3
EDOTs) is hypothesised to be related to coordination of
anions with the triblock copolymer within the VPP PEDOT.16
In direct comparison with PhenylPO22 –
4 , the insertion of Cl
22
yields a similar doping level to PhenylPO4 but a conductiv-
ity 2.5 times higher. Adding to this discussion is a compari-
son of the UV–vis-NIR spectra [Fig. 1(d) and Supporting
Information Figure S3] for the ratio of absorption at
1200 nm compared with 850 nm. The observed ratio for
ClO–4 and NO–3 in PEDOT is greater than the doping with
PhenylPO22 4 . Assigning this difference in ratio to changes in
the levels of bipolarons and polarons present is ambiguous,
given the overlap of their respective absorption (Supporting
Information Figure S1). Understanding this variation is of
further scientific research. However, this highlights that the
specific nature of the anion influences the resultant VPP
PEDOT properties, beyond those derived from different dop-
ing levels.
To investigate the relation of doping level and mobility with
conductivity, a broader range of PEDOT samples were tested
(chemically prepared and VPP). In the case of the VPP
PEDOT samples, a different PEG-PPG-PEG copolymer was
employed, and each sample respectively ion-exchanged with
the aforementioned anions. The different fabrication (primar-
ily owing to morphology changes as per ref. 8) and/or dop-
ing anions leads to markedly different electrical conductivity
values. As shown in Figure 1(e) the number of charge car-
riers is relatively constant across these samples. The open
circle data point represents an electrochemically reduced
VPP PEDOT variant to demonstrate that deliberate removal
of anions from PEDOT results in decreased numbers of
charge carriers. The doping level from XPS (fine scan spectra
shown in Supporting Information Figure S2) can be used to
determine the approximate number of charge carriers per
unit volume,11 and by assuming no contribution of electrons
to the charge transport the mobility of the holes can be cal-
culated from the Hall effect measurements. In Figure 1(f) the
order of magnitude increase in the mobility across these
samples correlates with the respective order of magnitude
increase in conductivity. In addition to Hall effect measure-
ments combined with XPS to determine this relationship,
THz spectroscopy was also employed to arrive at indepen-
dent measurements of the optical conductivity, charge carrier
 JOURNAL OF
POLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPER

FIGURE 2 Structural analysis of mixed doped PEDOT at high doping levels. 2D grazing incident wide angle x-ray scattering spectra
for (a) VPP PEDOT:Tos at high doping level after ion exchanged with Tos–, and ion exchanged with (e) ClO–4 and (f) PO–4 for com-
parison. For the spectra in (a), the integrations (b) along the surface normal showing the n00 lamellar peaks, (c) in the substrate
plane, showing the 020 p-stacking peak, and (d) along the radial through the mixed index reflection are presented. (g,h) collate the
integrations along the surface normal and substrate plane respectively, normalized to directly compare the anions studied herein.
[Color figure can be viewed at wileyonlinelibrary.com]

density, and mobility (see Supporting Information Table S2,
and as previously applied to PEDOT:PSS18). Figure 1(f)
shows that the relationship between the optically derived
transport properties overlay well with the electronically
derived transport properties.
2D GIWAXS analysis (Fig. 2 and Supporting Information Fig.
S4) provides information about the change in PEDOT chain
ordering for the type of anion electrochemically inserted into
the VPP PEDOT. All the samples exhibit pronounced crystal-
linity and are highly textured, with an “edge-on” orientation,

JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104 101

FULL PAPER
TABLE 1 The Determined Q Values and Corresponding d-Spacing for the PEDOT Structures Obtained using the Different Ion-
Exchanged Anions
[100] [010]
Anion d-Spacing (Å) d-Spacing (Å)
Tos 14.3 3.5
Cl 13.7 3.6
ClO4 13.1 3.6
NO3 12.6 3.5
PhenylPO4 13.7 3.6
that is, with lamellar stacking along the surface normal [Fig.
2(b)], and the p-stacking in the substrate plane [at Qxy 1.75
Å21 in Fig. 2(c)]. The peak widths of the 100 (lamellar) and
020 (p-stack) reflections correspond to domain sizes of 5.5
and 2.5 nm along the surface normal and in the substrate
plane, respectively, according to the Scherrer equation. The
presence of “mixed-index” reflections at (Qxy, Qz) 5 (1.0 Å21,
0.6 Å21) in Figure 2(d) indicates that 3D structures are
formed. All the ion exchanged samples exhibit even better
ordering relative to as-prepared VPP PEDOT:Tos.
There is a correlation between the type of anion inserted
into the VPP PEDOT and the lamellar spacing d. The d-spac-
ing changes from 14.0 Å for the Tos– samples to 12.6 Å for
the ClO–4 and NO–3 counter-ions (see Table 1). The other
anions of Cl– and PhenylPO22 –
4 are more similar to Tos (cor-
roborating prior studies of ion exchange of Tos with Cl– and
–
vice versa,38 and recent studies of VPP PEDOT:Tos treated
with NaOH and HCl39). The anion influence on the structure
correlates with the increased charge carrier mobility and
electrical conductivity for the ClO–4 and NO–3 samples and
varying levels of decreased conductivity for Cl– and
PhenylPO224 .
MD simulations and DFT calculation were conducted to
understand the role of the anion on PEDOT ordering and
electronic structure. Firstly, it is important to note that due
limitations of the calculations and simplification of the com-
putational model, they are unlikely to describe all the subtle
features seen in the experimental observations. Secondly, the
DFT calculations for the molecular orbitals (MOs) and DOSs
for a single chain with one anion (polaron) or two chains in
a p-stacking orientation with two anions (bipolaron) show
no differences between different counterions (Supporting
Information Figure S5 and S6). That is to say, the type of
anion has no influence on the electronic band structure of
the PEDOT when the geometric arrangement is held
constant.
However, MD simulations for the highly doped PEDOT show
that the morphology of the systems at hand strongly
depends on the type of anions and the doping level. The sim-
ulations were conducted in absence of a hard substrate
(which possibly induces strong texturing in “edge-on” orien-
tation). This is a likely scenario for the VPP process, where
nanofilms of polymer are formed at the liquid-vapor
102 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104
JOURNAL OF
WWW.POLYMERPHYSICS.ORG POLYMER SCIENCE
mixed [100] [010] mixed
d-Spacing (Å) Qz (1/Å) Qxy (1/Å) Qxy (1/Å) Qz (1/Å)
5.2 0.44 1.82 1.07 0.57
5.9 0.46 1.75 0.95 0.5
5.3 0.48 1.73 0.99 0.65
5.3 0.5 1.81 1 0.62
5.5 0.46 1.76 0.95 0.63
interface many micrometers away from the hard substrate.
Figure 3(a) shows representative snapshots of the PEDOT
structure illustrating formation of crystallites embedded in
an amorphous matrix. The calculated X-ray diffraction pat-
tern for all anions, Figure 3(e), shows a broad main peak at
Q 5 1.75 Å21 confirming the formation of crystallites of the
size of 14–16 Å with the interchain distance d 5 3.45 Å.
These values are in reasonable agreement with the experi-
mental GIWAXS measurements [Fig. 2(g,h) and Supporting
Information S4), in line with the widely reported packing
structure for PEDOT.
The structure of crystallites observed in the MD simulation
is, however, strikingly different depending on the anion type
and charge concentration. For the case of not too high dop-
ing level, 50% (corresponding to the cases of Tos–/Cl–,
Tos–/PO–4 ), the crystallites are composed of p-p stacked
PEDOT chains, with anions distributed randomly around the
crystallites [Fig. 3(b)]. For the case of very high doping lev-
els,  50% (corresponding to the cases of Tos–, Tos–/NO–3 ,
and Tos–/ClO–4 ) the morphology of crystallites undergoes
transformation with doping anions intercalating between
PEDOT chains to form a “sandwich” structure to screen the
positive charges in PEDOT chains. Interestingly, only planar
anions (Tos– and NO–3 ) intercalate between PEDOT chains
[Fig. 3(c)] whereas tetrahedral-shaped anions ClO–4 cannot
penetrate the space between PEDOT chains and thus stay
outside crystallites [Fig. 3(d)]. The formation of the interca-
lated structure is reflected in the radial distribution function
gP-P(r) for the carbon atoms describing the distance between
PEDOT chains, see Figure 3(e). For the cases of Tos–/Cl–,
Tos–/PhenylPO22 4 (when no intercalating structure forms) gP-
P(r) shows peaks at integer values of r/rp-p where the peak
r/rp-p 5 1 corresponds to the p-p distance between PEDOT
chains, rp-p 5 3.45 Å. For the remaining systems (Tos–, Tos–/
NO–3 , and Tos–/ClO–4 ) where the intercalating structure forms,
the peak at r/rp-p 5 1 is absent and the first peak in gP-P(r)
corresponds to r/rp-p 5 2. Note that for the case of Tos–/ClO–4
only Tos counterions intercalate between the PEDOT chains.
A signature of the intercalated structure is seen in the pre-
dicted X-ray scattering [Fig. 3(d)] at Q 5 0.95 Å21 corre-
sponding to the double p-p stacking distance 2rp-p 5 7 Å.
Note that the shoulder at Q 5 0.95 Å21 is apparently not
seen for either Cl– or PhenylPO22 4 . Figure 3(g,h) show dis-
tance distribution between the anions and PEDOT chains.
 JOURNAL OF
POLYMER SCIENCE WWW.POLYMERPHYSICS.ORG FULL PAPER

FIGURE 3 MD simulation of the PEDOT chain ordering for mixed anions at high doping levels. (a) MD Snapshot the PEDOT: Tos–/
NO–3 structures. PEDOT is shown in blue, Tos– is green, Cl– and NO–3 in red [water molecules, H atoms from PEDOT and Tos– and
O atoms and methyl groups from Tos– are not shown for clarity in (a–d)]. The snap-shot corresponds to a zoomed view of a repre-
sentative region of a computational box containing PEDOT crystallite, where a direction of a view is chosen perpendicular to the
crystallite to clearly see the intercalation effect. (b–d) Rendered images of PEDOT crystallites for the cases of Tos–/Cl–, Tos–/NO–3 ,
and Tos–/ClO–4 counterions, respectively. Stacking distances rp-p 5 3.45 Å and 2rp-p are indicated. “a” and “b” indicate counterions
contributing to corresponding peaks in (g and h). (c and d) correspond to the intercalated (“sandwich”) morphology. (e) X-ray Dif-
fraction Patterns for different counterions. Radial distribution functions: (f) gP-P(r) for the distance between PEDOT chains. (g) gS-
– – –
S(r) for the distance between SO3 from Tos and Sulfur from PEDOT. (h) gX -S(r) for the distance between center of mass of anion
and Sulfur from PEDOT. For all figures water content is 14% w/w. [Color figure can be viewed at wileyonlinelibrary.com]

For all anions, except Cl– and PhenylPO22 4 this distribution
shows similar features and is peaked at the distances 4
and 8 Å (peaks b and a, respectively). These peaks corre-
spond to anions situated on the side of chains as well as
between (or above/below) the chains, see Figure 3(b–d) for
illustration. These features cannot be unambiguously
assigned as an effect of intercalation in GIWAXS patterns,
however. GIWAXS simulation of the highly crystalline struc-
ture reported by Kim and Bredas40 also reveals features at
Qxy 5 1.1 Å21 (Supporting Information Figure S7). Further
structural investigation is necessary to validate the hypothe-
sis arising from the MD simulations at high doping levels
of  50%.
CONCLUSIONS
In summary, the doping anion plays a critical role in defining
the electronic transport properties of PEDOT. When ClO–4 is
introduced into the structure to form PEDOT:Tos/ClO4 the
electrical conductivity approaches 3000 S/cm with a mobility
of > 3 cm2/Vs. While the molecular orbitals and nature of
the charge carriers in PEDOT chains do not alter, at high
doping levels the specific anion chosen leads to changes in
the morphological structure of PEDOT. GIWAXS indicates
highly ordered PEDOT chains, while MD simulations give
insight that certain anions may actually intercalate into the
p-p stacking of PEDOT. Enhancement of the charge carrier
mobility towards that of metals, as well as the crystallinity,
are critical to realizing the uptake of conducting polymers
such as PEDOT in low-cost, flexible electronic devices.
ACKNOWLEDGMENTS
The authors acknowledge the European Research Council
(ERC-starting-grant 307596), the Swedish foundation for stra-
tegic research (project: “Nano-material and Scalable TE mate-
rials”), the Knut and Alice Wallenberg foundation (project
“Power Paper” and “Tail of the Sun”), The Swedish Energy
Agency (38332 2 1), the Swedish Research Council via
“Research Environment grant” on “Disposable paper fuel cells”
(2016 2 05990), and the Advanced Functional Materials Cen-
ter at Link€ oping University. The authors thank Masatsugu
Yamashita from the THz Sensing & Imaging Lab at RIKEN in
Japan for conducting the THz reflectance spectroscopy experi-
ments. D.R. Evans acknowledges the support of the Australian
Research Council through the Future Fellowship scheme
(FT160100300). J.W. Andreasen acknowledges the support of
the European Research Council (ERC-consolidator-grant
681881). The computations were performed on resources pro-
vided by the Swedish National Infrastructure for Computing
(SNIC) at NSC.
REFERENCES AND NOTES
1 C. Chiang, C. Fincher, Jr., Y. Park, A. Heeger, H. Shirakawa, E.
Louis, S. Gau, A. G. MacDiarmid, Phys. Rev. Lett. 1977, 39, 1098.
2 Y. Cao, G. M. Treacy, P. Smith, A. J. Heeger, Appl. Phys. Lett.
1992, 60, 2711.
3 S. J. Pomfret, P. N. Adams, N. P. Comfort, A. P. Monkman,
Adv. Mater. 1998, 10, 1351.
4 M. Yamaura, T. Hagiwara, K. Iwata, Synth. Met. 1988, 26,
209.

JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104 103

FULL PAPER WWW.POLYMERPHYSICS.ORG
5 (a) M. Fabretto, C. Jariego-Moncunill, J. P. Autere, A.
Michelmore, R. D. Short, P. Murphy, Polymer. 2011, 52, 1725;
(b) D. C. Martin, J. Wu, C. M. Shaw, Z. King, S. A. Spanninga,
S. Richardson-Burns, J. Hendricks, J. Yang, Polym. Rev. 2010,
50, 340.
6 C. Badre, L. Marquant, A. M. Alsayed, L. A. Hough, Adv.
Func. Mater. 2012, 22, 2723.
7 Y. Xia, K. Sun, J. Ouyang, Adv. Mater. 2012, 24, 2436.
8 M. V. Fabretto, D. R. Evans, M. Mueller, K. Zuber, P. Hojati-
Talemi, R. D. Short, G. G. Wallace, P. J. Murphy, Chem. Mater.
2012, 24, 3998.
9 M. N. Gueye, A. Carella, N. Massonnet, E. Yvenou, S. Brenet,
J. Faure-Vincent, S. Pouget, F. Rieutord, H. Okuno, A. Benayad,
Chem. Mater. 2016, 28, 3462.
10 (a) Y. Li, Y. Zou, Adv. Mater. 2008, 20, 2952; (b) I.
McCulloch, M. Heeney, C. Bailey, K. Genevicius, I. MacDonald,
M. Shkunov, D. Sparrowe, S. Tierney, R. Wagner, W. Zhang,
Nat. Mater. 2006, 5, 328.
11 B. Cho, K. S. Park, J. Baek, H. S. Oh, Y.-E. Koo Lee, M. M.
Sung, Nano Lett. 2014, 14, 3321.
12 O. Bubnova, Z. U. Khan, H. Wang, S. Braun, D. Evans, M.
Fabretto, P. Hojati-Talemi, D. Dagnelund, J.-B. Arlin, Y. Geerts,
S. Desbief, D. Breiby, J. W. Andreasen, R. Lazzaroni, W. Chen,
I. Zozoulenko, M. Fahlman, P. Murphy, M. Berggren, X. Crispin,
Nat. Mater. 2014, 13, 190.
13 (a) O. Bubnova, Z. U. Khan, A. Malti, S. Braun, M. Fahlman, M.
Berggren, X. Crispin, Nat. Mater. 2011, 10, 429; (b) T. Park, C.
Park, B. Kim, H. Shin, E. Kim, Energy Environ. Sci. 2013, 6, 788.
14 K. Ando, S. Watanabe, S. Mooser, E. Saitoh, H. Sirringhaus,
Nat. Mater. 2013, 12, 622.
15 (a) B. Winther-Jensen, O. Winther-Jensen, M. Forsyth, D. R.
MacFarlane, Science. 2008, 321, 671; (b) P. Cottis, D. Evans, M.
Fabretto, S. Pering, P. J. Murphy, P. Hojati-Talemi, RSC Adv.
2014, 4, 9819.
16 C. Karlsson, J. Nicholas, D. Evans, M. Forsyth, M. Strømme,
M. Sjo€ din, P. C. Howlett, C. Pozo-Gonzalo, ChemSusChem.
2016, 9, 2112.
17 J. F. Franco-Gonzalez, I. V. Zozoulenko, J. Phys. Chem. B.
2017, 121, 4299.
18 M. Yamashita, C. Otani, M. Shimizu, H. Okuzaki, Appl. Phys.
Lett. 2011, 99, 143307.
19 D. Apitz, R. Bertram, N. Benter, W. Hieringer, J. W.
Andreasen, M. M. Nielsen, P. Johansen, K. Buse, Phys. Rev. E.
2005, 72, 036610.
20 J. Wang, R. M. Wolf, J. W. Caldwell, P. A. Kollman, D. A.
Case, J. Comput. Chem. 2004, 25, 1157.
104 JOURNAL OF POLYMER SCIENCE, PART B: POLYMER PHYSICS 2018, 56, 97–104
JOURNAL OF
POLYMER SCIENCE
21 A. I. Jewett, Z. Zhuang, J.-E. Shea, Biophys. J. 2013, 104,
169a.
22 S. Plimpton, J. Comput. Phys. 1995, 117, 1.
23 H. Berendsen, J. Postma, W. Van Gunsteren, J. Hermans, In
The Jerusalem Symposia on Quantum Chemistry and Bio-
chemistry, B. Pullman, Ed.; Springer: Dordrecht, April 1981;
Vol. 14, pp 13.
24 (a) A. Ugur, F. Katmis, M. Li, L. Wu, Y. Zhu, K. K. Varanasi,
K. K. Gleason, Adv. Mater. 2015, 27, 4604; (b) A. Elschner, S.
Kirchmeyer, W. Lovenich, U. Merker, K. Reuter, PEDOT: Princi-
ples and applications of an intrinsically conductive polymer;
CRC Press: Florida, 2010.
25 T. Yokozawa, A. Yokoyama, Chem. Rev. 2009, 109, 5595.
26 (a) W. Shinoda, M. Shiga, M. Mikami, Phys. Rev. B. 2004,
69, 134103; (b) G. J. Martyna, D. J. Tobias, M. L. Klein, J.
Chem. Phys. 1994, 101, 4177; (c) M. Parrinello, A. Rahman, J.
Appl. Phys. 1981, 52, 7182.
27 M. E. Tuckerman, J. Alejandre, R. Lo  pez-Rendo n, A. L.
Jochim, G. J. Martyna, J. Phys. A: Math. Gen. 2006, 39, 5629.
28 S. Coleman, D. Spearot, L. Capolungo, Modell. Simul.
Mater. Sci. Eng. 2013, 21, 055020.
29 Y.-S. Lin, G.-D. Li, S.-P. Mao, J.-D. Chai, J. Chem. Theory
Comput. 2012, 9, 263.
30 R. Krishnan, J. S. Binkley, R. Seeger, J. A. Pople, J. Chem.
Phys. 1980, 72, 650.
31 U. C. Singh, P. A. Kollman, J. Comput. Chem. 1984, 5, 129.
32 J. L. Bredas, G. B. Street, Acc. Chem. Res. 1985, 18, 309.
33 J. J. Apperloo, L. B. Groenendaal, H. Verheyen, M.
Jayakannan, R. A. J. Janssen, A. Dkhissi, D. Beljonne, R.
das, Chem. Eur. J. 2002, 8, 2384.
Lazzaroni, J.-L. Bre
34 W. A. Munoz, S. K. Singh, J. Franco-Gonzalez, M. Linares,
X. Crispin, I. Zozoulenko, Phys. Rev. B. 2016, 94, 205202.
35 B. Winther-Jensen, M. Forsyth, K. West, J. W. Andreasen, P.
Bayley, S. Pas, D. R. MacFarlane, Polymer 2008, 49, 481.
36 G. Zotti, S. Zecchin, G. Schiavon, F. Louwet, L.
Groenendaal, X. Crispin, W. Osikowicz, W. Salaneck, M.
Fahlman, Macromolecules 2003, 36, 3337.
37 P.-H. Aubert, L. Groenendaal, F. Louwet, L. Lutsen, D.
Vanderzande, G. Zotti, Syn. Met. 2002, 126, 193.
38 B. Winther-Jensen, M. Forsyth, K. West, J. W. Andreasen,
G. G. Wallace, D. R. MacFarlane, Org. Electron. 2007, 8, 796.
39 Z. U. Khan, O. Bubnova, M. J. Jafari, R. Brooke, X. Liu, R.
Gabrielsson, T. Ederth, D. R. Evans, J. W. Andreasen, M.
Fahlman, J. Mater. Chem. C. 2015, 3, 10616.
das, J.
40 E. G. Kim, J. L. Bre Am. Chem. Soc. 2008, 130, 16880.